LLMs are not (consistently) Bayesian: Quantifying internal (in)consistencies of LLMs' probabilistic beliefs

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Chen, Chacha, Jörke, Matthew, Goliński, Adam, Fedzechkina, Masha, Sapiro, Guillermo, Williamson, Sinead, Foti, Nicholas
Format:	Preprint
Published:	2026
Subjects:	Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2605.06915